220 Urbanization models and the regions
results from region to region. Thus the satellite-cities model is consistently the
best among regions for 10 of the 18 attributes (biodiversity sites, recreation/
tourism sites, nearby slopes facing cities, etc.). The dispersed-sites model, on
theother hand, is consistently the worst among urban regions for seven of the
attributes.
Is there a single attribute, or two or three, that produces results similar to that
forthe whole set of 18 variables? Rivers/major streams and reservoirs/lakes are
theonly two attributes with the same best-to-worst ( 1 , 2 , 3 ,and 4 )model results
as for the whole set (Table 8.2). The actual averages, as well as the consistency
in results across regions, for therivers/streams attributeare closest to those of the
total set.Measuring the length of rivers and major streams affected or degraded
bydevelopment may be a particularly useful measure to differentiate mosaic-
sequence models and to evaluate the overall effects of urbanization.
Assaying a global set of cities with 18 informative variables may seem ambi-
tious, even sufficient. Yet other variables or considerations may be important for
the models as the following cases illustrate.Concentric-zones model: additional dimensions
Expanding outward from a metropolitan area would likely increase the
heat-island effect for cities where that is a problem. The concentric-zones trajec-
tory would also be environmentally detrimental for nearby slopes facing a city,
and would leave city center residents increasingly isolated from greenspace of
theurban-region ring (Table 8.2). Models that eliminate greenspace wedges in
ametro area, or alternatively, continue green wedges outward as urbanization
proceeds, would be interesting.
Abulges model instead of the concentric-zones model might produce lit-
tle difference in response. However, if the bulges were on especially valuable
greenspace, the concentric model would be better. Conversely, targeting devel-
opment onto greenspace of low value should make the bulges approach even
better than the concentric-zones model.Satellite-cities model: additional dimensions
Concentrating growth around satellite cities is a dispersal process that
may create new major cities in the outer urban-region ring where natural sys-
tems are most valuable. Or several rather than a few satellites could be nuclei for
urbanization. This might create a pattern somewhat akin to dispersed sites, with
arelatively extensive associated road network. The satellite-cities approach would
normally be the best of the four models for maintaining greenspace wedges, a
greenbelt, or an urban growth boundary.Three plan options for the region
expected future development around five satellite cities (Color Figure6)(Forman